Both functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI) can be helpful in characterizing the neurophysiological status of patients with aphasia. Such measurements can be used to develop treatment approaches, select patient-specific treatments, and gauge the effectiveness of treatment. However, brain imaging in aphasia must be improved. Both fMRI and DTI are highly sensitive to both head motion and oropharyngeal motion, and these artifacts are exacerbated in patients with non-fluent aphasia, who have a more difficult time producing speech. Stroke is often precipitated by and/or accompanied by alterations in vasculature, which can affect the MRI signal. Hemodynamic redistribution leads to ambiguity about the source of MRI signals presumed to indicate neural reorganization and/or remodeling. Artifacts from reperfusion hemorrhage leads to susceptibility artifact in the vicinity of lesions. The proposed work will specifically address each of these technical issues to facilitate the use of neurophysiological (functional imaging) measures in patients with aphasia. For fMRI and DTI to guide therapeutic decision-making, there must be formal ways to characterize the physiological state of individual patients and to correlate biological and behavioral parameters. In this proposal, we develop two classes of measures, scalar indices and network activation maps, and investigate their relationship to measures of speech and language. Scalar indices reduce an entire pattern of activation into a single number, such as degree of brain lateralization or integrity of callosal white matter. Network maps include structural equation models and Euclidean distance maps, and we develop methods to compare such large-scale models between individuals or groups (e.g., patients who do and do not benefit from therapy). This component of the research aims to develop rational neurophysiological measures for aphasia research. Finally, these methodological improvements will be applied to a treatment study for patients with chronic non-fluent aphasia and left frontal lesions. The therapeutic intervention is based on (a) neurobiological observations that the brain has circuits particularly active in motor imitation through observation-execution matching, including oral motor imitation; (b) behavioral observations that imitation is a cornerstone of many treatments for non-fluent aphasia; and (c) treatment data that favor intensive approaches and graded incremental learning. Outcomes are assessed by both physiological and behavioral measures, which are collected every three weeks starting six weeks before treatment and ending six weeks after completion of treatment.